Georg Hansmann, MD; Roger A. Wagner, MD, PhD; Stefan Schellong, BA;
Vinicio A. de Jesus Perez, MD; Takashi Urashima, MD; Lingli Wang, MD;
Ahmad Y. Sheikh, MD; Renée S. Suen, BSc; Duncan J. Stewart, MD; Marlene Rabinovitch, MD
Background—Patients with pulmonary arterial hypertension (PAH) have reduced expression of apolipoprotein E (apoE) and peroxisome proliferator–activated receptor-␥ in lung tissues, and deficiency of both has been linked to insulin resistance. ApoE deficiency leads to enhanced platelet-derived growth factor signaling, which is important in the pathobiology of PAH. We therefore hypothesized that insulin-resistant apoE-deficient (apoE⫺/⫺) mice would develop PAH that could be reversed by a peroxisome proliferator–activated receptor-␥agonist (eg, rosiglitazone).
Methods and Results—We report that apoE⫺/⫺mice on a high-fat diet develop PAH as judged by elevated right ventricular systolic pressure. Compared with females, male apoE⫺/⫺were insulin resistant, had lower plasma adiponectin, and had higher right ventricular systolic pressure associated with right ventricular hypertrophy and increased peripheral pulmonary artery muscularization. Because male apoE⫺/⫺mice were insulin resistant and had more severe PAH than female apoE⫺/⫺mice, we treated them with rosiglitazone for 4 and 10 weeks. This treatment resulted in markedly higher plasma adiponectin, improved insulin sensitivity, and complete regression of PAH, right ventricular hypertrophy, and abnormal pulmonary artery muscularization in male apoE⫺/⫺ mice. We further show that recombinant apoE and adiponectin suppress platelet-derived growth factor-BB–mediated proliferation of pulmonary artery smooth muscle cells harvested from apoE⫺/⫺or C57Bl/6 control mice.
Conclusions—We have shown that insulin resistance, low plasma adiponectin levels, and deficiency of apoE may be risk factors for PAH and that peroxisome proliferator–activated receptor-␥activation can reverse PAH in an animal model.
(Circulation. 2007;115:1275-1284.)
Key Words:apolipoproteins 䡲 glucose 䡲 hypercholesterolemia 䡲 hypertension, pulmonary 䡲 insulin 䡲metabolism 䡲 PPAR gamma
A
lthough insulin resistance is associated with systemic cardiovascular disease,1–3 it has not been implicated as a predisposing factor in pulmonary arterial hypertension (PAH).Several findings, however, support such an association. Patients with idiopathic PAH have reduced pulmonary mRNA expres-sion of peroxisome proliferator–activated receptor gamma (PPAR␥),4a ligand-activated nuclear receptor and transcription factor that regulates adipogenesis and glucose metabolism.5–7 They also have reduced pulmonary mRNA expression of apo-lipoprotein E (apoE),8 a protective factor known to reduce circulating oxidized low-density lipoprotein and atherogenesis in the vessel wall.9Deficiency of both PPAR␥and apoE has been linked to insulin resistance and the metabolic syndrome.7,9 Elevated levels of several circulating factors that are normally
repressed by PPAR␥are associated with insulin resistance3and implicated in the pathobiology of PAH. These include interleukin-6,10,11 fractalkine,12,13 monocyte chemoattractant protein-1,14endothelin-1 (ET-1),15–17and the endogenous nitric oxide synthase inhibitor asymmetric dimethylarginine (ADMA).18,19
Clinical Perspective p 1284
Heightened signaling by platelet-derived growth factor-BB (PDGF-BB)/mitogen-activated protein kinase leading to smooth muscle cell (SMC) proliferation and migration is also a key clinical feature of pulmonary vascular disease.20 –22 With apoE deficiency, abundant oxidized low-density li-poprotein23and PDGF-BB23were shown to induce
mitogen-Received September 8, 2006; accepted December 29, 2006.
From the Department of Pediatrics, Division of Pediatric Cardiology (G.H., S.S., V.A.D.J.P., T.U., L.W., M.R.), Department of Medicine, Division of Cardiovascular Medicine (R.A.W.), and Department of Cardiovascular Surgery (A.Y.S.), Stanford University School of Medicine, Stanford, Calif, and Department of Medicine, Division of Cardiology, University of Toronto, Toronto, Ontario, Canada (R.S.S., D.J.S.).
The online-only Data Supplement, consisting of expanded Methods and tables, is available with this article at http://circ.ahajournals.org/cgi/content/full/
CIRCULATIONAHA.106.663120/DC1.
Correspondence to Dr Marlene Rabinovitch, Vera Moulton Wall Center for Pulmonary Vascular Disease, Stanford University School of Medicine, CCSR 2245B, 269 Campus Dr, Stanford, CA 94305-5162. E-mail marlener@stanford.edu
© 2007 American Heart Association, Inc.
Circulationis available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.106.663120
genes (eg, cyclin D1), and subsequently proliferation and migration of systemic vascular SMCs.23,24Interestingly, high glucose concentrations induce mitogen-activated protein kinase/phosphatidylinositol 3-kinase (PI3K)– dependent up-regulation of PDGF receptor- (PDGFR-) and potentiate SMC migration in response to PDGF-BB.24In concert with PI3K, PDGFR-/mitogen-activated protein kinase-signaling also leads to SMC resistance to apoptosis.25
In systemic vascular SMCs, apoE and adiponectin,26 a PPAR␥ target in adipocytes,7 inhibit PDGF-BB–induced SMC proliferation and migration.23,27ApoE internalizes the PDGFR-,28 –30and adiponectin sequesters the ligand PDGF-BB.31 Thus, in association with insulin resistance, reduced levels of apoE23and adiponectin27can be expected to enhance PDGF-BB signaling. In accordance with these observations, diabetic apoE-deficient (apoE⫺/⫺) mice show pronounced PDGF-BB signaling and neointimal thickening of the arterial vessel wall.32We hypothesized that these factors would have similar effects on pulmonary arterial SMCs (PASMCs), consequently leading to PAH.
PPAR␥agonists are clinically used to make cells insulin sensitive, thereby obviating the detrimental effects of insulin resistance related to hyperlipidemia, inflammation, and mito-genesis in the vessel wall.33Rosiglitazone, a PPAR␥ligand of the thiazolidinedione class, enhances insulin-mediated glu-cose uptake and inhibits proliferation and migration of systemic SMCs induced by PDGF-BB.33,34 We therefore hypothesized that PAH would develop in apoE⫺/⫺ insulin-re-sistant mice but that the disease process would be attenuated or reversed by a PPAR␥agonist (eg, rosiglitazone).
In the present study, we show that apoE deficiency, in association with a high-fat (HF) diet, leads to both insulin resistance and PAH. Male apoE⫺/⫺ mice had more severe PAH (ie, higher right ventricular systolic pressure [RVSP], right ventricular hypertrophy [RVH], and enhanced periph-eral PA muscularization) associated with insulin resistance and lower plasma adiponectin levels compared with female apoE⫺/⫺mice. Because testosterone inhibits the secretion of adiponectin in adipocytes,35we hypothesized that elevation of this vasoprotective adipocytokine may account for the less severe vascular phenotype in female apoE⫺/⫺ mice. We therefore treated male apoE⫺/⫺ mice with rosiglitazone and documented 8-fold-higher plasma adiponectin levels, im-proved insulin sensitivity, and complete regression of PAH, RVH, and abnormal PA muscularization. To further establish a direct link between apoE and adiponectin and SMC prolif-eration and survival, we treated murine (apoE⫺/⫺and wild-type) PASMCs in culture with recombinant apoE and adi-ponectin. We showed that both proteins inhibit PDGF-BB–
induced proliferation in apoE⫺/⫺and wild-type PASMCs. Our data therefore suggest that insulin resistance and deficiency of apoE and/or adiponectin may be risk factors for PAH that can be reversed by PPAR␥activation.
Methods
Expanded Methods and Results sections are given in the online Data Supplement.
ApoE mice (B6.129P2-Apoetm1Unc/J) and C57Bl/6 control mice were obtained from Jackson Laboratories (Bar Harbor, Me). At 4 weeks of age, the mice were either continued on regular chow or switched to HF diet (Dyets No. 101511, Dyets Inc, Bethlehem, Pa) for a maximum of 21 weeks. For the nontreatment study, 15-week-old male and female mice (apoE⫺/⫺, C57Bl/6 controls) on either diet were studied. For the rosiglitazone treatment study, 15-week-old male mice (apoE⫺/⫺, C57Bl/6 controls) on HF diet were used. Half of the animals received rosiglitazone (GlaxoSmithKline, Research Tri-angle Park, NC) 10 mg/kg body weight per day PO incorporated into the food for 4 or 10 weeks. All protocols were approved by the Stanford Animal Care Committee.
Hemodynamic Measurements
Measurements of RVSP and RV dP/dt were performed by jugular vein catheterization (1.4F, Millar Instruments Inc, Houston, Tex) under isoflurane anesthesia (1.5% to 2.5%) using a closed-chest technique in unventilated mice at 15, 19, and 25 weeks of age. Left ventricular (LV) end-diastolic pressure was determined by LV catheterization via the left carotid artery under isoflurane anesthesia.
Systemic blood pressure was determined by the tail-cuff method in nonanesthesized mice. Measurements of cardiac output and function were performed by echocardiography.
RVH and LV Hypertrophy
RVH was measured by the weight of the RV relative to LV⫹septum.
LV hypertrophy was measured as absolute weight of the LV plus septum. LV dilatation was assessed by echocardiographic M-mode measurement of the LV end-diastolic inner diameter.
Lung Tissue Preparation
Lungs were perfused with normal saline, fixed in 10% formalin overnight, and then either embedded in paraffin for standard histol-ogy or frozen for Oil-red-O staining. A subset of left lungs (approximately half) were barium infused via PA-inserted tubing to label peripheral PAs for morphometric analysis and micro– com-puted tomography (CT) imaging.
Morphometric Analysis
Transverse left lung sections were stained by elastic van Gieson and Movat pentachrome stains. From all mice, we took the same full section in the mid portion of the lung parallel to the hilum and embedded it in the same manner. Muscularization was assessed in barium-injected left lung sections by calculating the proportion of fully and partially muscularized peripheral (alveolar wall) PAs to total peripheral PAs. All measurements were done blinded to genotype and condition.
Micro-CT Imaging
A custom-built eXplore Locus RS120 Micro CT Scanner (GE Health Care, Ontario, Canada) was used to acquire nondestructing 3-dimensional images of barium-infused whole-lung specimens. Images were scanned at 49-m resolution and 720 views (70 kV [peak], 50 mAmps, 30-ms single image acquisition time) and reconstructed with the eXplore Reconstruction Utility, and volumes were viewed and rendered with the GE Health Care MicroView software.
Fasting Whole-Blood and Plasma Measurements
We performed tail vein puncture in nonanesthetized, overnight-starved mice, followed by duplicate whole-blood glucose measure-ments with a glucometer (Freestyle/Abbott). Fasting blood plasma was obtained via retro-orbital bleeding or cardiac puncture. White blood cell count and hematocrit were assessed by the Stanford Animal Facility Laboratories. Hemoglobin A1c was measured by Esoterix (Calabasa Hills, Calif). Plasma ET-1 was measured by ELISA. Plasma ADMA levels were determined by high-performance liquid tomography at Oxonon Bioanalysis Inc (Oakland, Calif). All
nostics (St Charles, Mo).
Cell Culture
Primary murine PASMCs were isolated from apoE⫺/⫺and C57Bl/6 mice using a modified elastase/collagenase digestion protocol as previously described.36Murine PASMCs were grown to 70% con-fluence and cultured in starvation media (DMEM, 0.1% FBS, 100 U/mL penicillin, 0.1 mg/mL streptomycin) for 24 hours. Recombi-nant apoE (Chemicon International, Temecula, Calif) and adiponec-tin (BioVision, Mountain View, Calif) were added to quiescent cells 30 minutes before mitogenic stimulation with PDGF-BB (R&D Systems, Minneapolis, Minn).
Cell Proliferation Assays
For cell counts, PASMCs were seeded at 2.5⫻104cells per well of a 24-well plate in growth medium and allowed to adhere overnight.
The medium was removed, and the cells were washed 3 times with PBS and incubated in starvation media for 24 hours, followed by PDGF-BB stimulation (20 ng/mL) for 0 and 72 hours. Cells were then washed with PBS, trypsinized, resuspended, and counted in a hemacytometer.
Statistical Analysis
Values from multiple experiments are expressed as mean⫾SEM.
Using the Kohmogorov-Smirnov test and larger data sets from previous studies, we could show that the measured values were approximately normally distributed. Statistical significance was determined using 1-way ANOVA, followed by Bonferroni’s multiple-comparison test unless stated otherwise. A value ofP⬍0.05 was considered significant. The significance of our data also was confirmed by the nonparametric Mann-Whitney test. The number in each group is indicated in the column graphs and in the figure legends. For some of the metabolic measurements such as blood glucose, which did not require invasive blood draws, a larger number of animals could be assessed, resulting in minor unevenness in the numbers reported.
All authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Results
ApoEⴚ/ⴚand C57Bl/6 Mice on Regular Chow Have Similar RVSPs and RV Mass
First, we assessed 15-week-old male and female apoE⫺/⫺
mice on regular chow for the presence and severity of PAH.
Values of RVSP, a measure of PAH, and of RV/LV⫹septum ratio, a measure of RVH, were similar in apoE⫺/⫺ and C57Bl/6 control mice. Moreover, no significant differences were observed in RVSP or RV/LV⫹septum ratio between genders of either genotype (Table 1).
ApoEⴚ/ⴚMice on HF Diet Develop PAH
An 11-week HF diet treatment did not significantly increase RVSP in 15-week old C57Bl/6 mice (Table 1). In contrast, apoE⫺/⫺ mice on HF diet for the same duration developed PAH as judged by significant elevation in RVSP, with males having higher values than females (Figure 1A and Table 1).
In addition, only male apoE⫺/⫺mice on HF diet had RVH and enhanced peripheral PA muscularization (Figure 1B and 1C) compared with C57Bl/6 controls (P⬍0.001). A direct com-parison revealed a more severe PAH phenotype in the male versus female apoE⫺/⫺mice on HF diet in that RVSP (28.9 versus 24.9 mm Hg; P⫽0.0014), RVH (RV/LV⫹septum ratio, 0.41 versus 0.29;P⫽0.0093), and peripheral
muscular-greater (Figure 1A through 1C, unpaired 2-tailedttest). RV systolic function (RV dP/dtmax) was augmented in apoE⫺/⫺
mice of both genders and reflected the elevated RVSP compared with C57Bl/6 controls (Table 1). RV diastolic function (RV dP/dtmin) was greater in both male (trend) and female (P⬍0.05) apoE⫺/⫺mice compared with C57Bl/6 mice of the same gender. Systemic blood pressure, cardiac output, LV function (indicated by LV end-diastolic pressure of 2 to 5 mm Hg), and hematocrit were similar in both genotypes (Table 1). Thus, the elevation of RVSP in apoE⫺/⫺compared with C57Bl/6 mice likely reflected an elevation in pulmonary vascular resistance.
PA Atherosclerosis in the ApoEⴚ/ⴚMice Does Not Cause Significant PA Stenosis
Further studies were carried out to determine whether apoE⫺/⫺mice on HF diet developed, in addition to neomus-cularization of peripheral PAs, occlusive atheroma account-ing for the elevated RVSP and RVH compared with C57Bl/6 mice. Micro-CT imaging of barium-injected lungs revealed an irregularly shaped main PA vessel wall in male and female apoE⫺/⫺mice on HF diet but excluded PA branch stenosis as contributing to the RVSP elevation (Figure 1F and 1G). This feature was associated with nonocclusive atherosclerotic lesions only in large intrapulmonary arteries (diameterⱖ500
m) in apoE⫺/⫺mice of both genders on HF diet (see Figure I in the online Data Supplement). Atherosclerotic lesions were neither seen in C57Bl/6 control mice on HF diet nor in mice of both genotypes on regular chow.
Insulin Resistance Is Associated With More Severe PAH in Male ApoEⴚ/ⴚMice
We focused our attention on possible differences in the lipid profile and markers of insulin resistance associated with the more severe PAH phenotype in apoE⫺/⫺versus C57Bl/6 mice, particularly in male versus female apoE⫺/⫺ mice. Higher plasma cholesterol (mainly non– high-density lipoprotein cholesterol) was observed in male and female apoE⫺/⫺mice compared with C57Bl/6 controls on HF diet. In addition, apoE⫺/⫺mice had moderately higher triglyceride levels that were similar in male and female apoE⫺/⫺mice and indepen-dent of the diet (Tables I and II in the online Data Supple-ment). However, only in male apoE⫺/⫺ and not female apoE⫺/⫺mice on HF diet did we observe features consistent with insulin resistance (ie, elevated fasting blood glucose and insulin levels), compared with C57Bl/6 controls (Figure 2).
Because of the concordance of insulin resistance and the more severe PAH phenotype in male versus female apoE⫺/⫺
mice, we investigated whether the females had higher plasma levels of the insulin-sensitizing adipocytokines adiponectin and leptin. Although the HF diet resulted in a marked increase in adiponectin and leptin levels in control C57Bl/6 mice of both genders, this marked upregulation was absent in apoE⫺/⫺
mice (for adiponectin, see Figure 2A and 2B; for leptin, see Tables I and II in the online Data Supplement). It is therefore possible that the release of adiponectin and leptin from adipocytes in response to HF diet is to some extent apoE dependent. However, female mice of both genotypes had
⬃50% higher plasma adiponectin levels than their male counterparts (Figure 2A and 2B). Gender differences in leptin levels were difficult to ascertain because of consid-erable variability in the individual values (Tables I and II in the online Data Supplement). Because testosterone inhibits the secretion of adiponectin in adipocytes,35 we hypothesized that higher adiponectin levels, in association with lack of insulin resistance, may account for the less severe pulmonary vascular phenotype in female apoE⫺/⫺
mice on a HF diet.
PPAR␥Activation Elevates Plasma Adiponectin, Improves Insulin Sensitivity, and Reverses PAH On the basis of these data, we reasoned that the presence of PAH may be determined by the inability to sufficiently raise adi-ponectin levels in association with a HF diet (Figure 2A and 2B) and that the severity of the disease may be a function of the
degree of hyperinsulinemia and hyperglycemia (Figure 2C through 2F). We therefore hypothesized that treating the pulmo-nary hypertensive, insulin-resistant male apoE⫺/⫺ mice with a PPAR␥ agonist to increase plasma adiponectin and improve insulin sensitivity might arrest disease progression or reverse PAH. Thus, we treated 15-week-old male apoE⫺/⫺and C57Bl/6 control mice on a HF diet with rosiglitazone 10 mg · kg⫺1· d⫺1 incorporated into their food. The effects of both 4- and 10-week treatments on RVSP, RVH, and metabolic features were as-sessed. Four-week treatment with rosiglitazone resulted in much higher plasma adiponectin levels in C57Bl/6 control (5-fold) and apoE⫺/⫺(8-fold) mice compared with untreated animals of the same genotype (Figure 3A). The higher plasma adiponectin in treated versus untreated apoE⫺/⫺mice was associated with lower blood glucose (control level), indicating improved insulin sen-sitivity in apoE⫺/⫺mice treated with rosiglitazone (Figure 3A, 3C, and 3E).
Echocardiographic, and Heart Weight Measurements in C57Bl/6 (Control) and ApoE Mice on HF Diet
Control Males ApoE⫺/⫺Males Control Females ApoE⫺/⫺Females P n
Mice on regular chow
RVSP, mm Hg 20.6⫾0.8 23.2⫾0.6 19.7⫾1.5 22.5⫾0.8 4–5
RV, mg 26.5⫾1.3‡ 24.9⫾1.4† 17.9⫾0.9 18.1⫾0.7 CM vs CF,‡ AM vs AF† 4–5
RV/LV⫹S 0.26⫾0.01 0.30⫾0.01 0.26⫾0.01 0.27⫾0.02 4–5
LV⫹S, mg 101.0⫾2.7‡ 84.6⫾4.3* 69.2⫾2.7 68.6⫾2.2 CM vs AM,* CM vs CF,‡ AM vs AF* 4–5
Mice on HF diet Hemodynamics
RVSP, mm Hg 20.6⫾0.5 28.9⫾0.6‡ 20.5⫾0.9 24.9⫾0.6† AM vs CM,‡ AM vs AF,† AF vs CF† 4–5
RV dP/dtmax, mm Hg/s 1132⫾82 1754⫾62* 950⫾206 1513⫾125* AM vs CM,* AF vs CF* 4–5
RV dP/dtmin, mm Hg/s ⫺951⫾86 ⫺1396⫾72 ⫺753⫾173 ⫺1279⫾96* AF vs CF* 4–5
Systolic BP, mm Hg 92⫾1.9 99⫾2.6 83⫾5.4 96⫾1.3 4–5
MAP, mm Hg 79⫾1.4 85⫾2.9 77⫾3.5 86⫾1.5 4–5
Diastolic BP, mm Hg 73⫾1.6 77⫾4.4 71⫾3.4 80⫾1.8 4–5
LVEDP, mm Hg 2.4⫾0.5 2.8⫾0.2 2.5⫾0.3 2.7⫾1.2 4–5
Echocardiography 4–5
Heart rate, bpm 398⫾29.6 370⫾22.5 447⫾27 408⫾36 4–5
EF, % 74⫾0.7 77.6⫾5.4 70.2⫾2.4 85.7⫾2.6* AF vs CF* 4–5
FS, % 37.5⫾0.6 42.0⫾6.2 34.3⫾1.9 51.2⫾2.9† AF vs CF† 4–5
CO, mL/min) 30.0⫾2.1 29.1⫾5.5 24.3⫾2.2 24.7⫾3.8 4–5
LVIDD, mm 3.4⫾0.06 3.4⫾0.28 3.1⫾0.02 3.0⫾0.13 4–5
LVISD, mm 2.1⫾0.02 2.0⫾0.33 2.0⫾0.06 1.5⫾0.16 4–5
Heart weight
RV, mg 19.5⫾1.3 33.0⫾4.0† 18.1⫾0.5 18.0⫾1.1 AM vs CM,† AM vs AF† 4
RV/LV⫹S 0.25⫾0.02 0.41⫾0.03‡ 0.24⫾0.01 0.29⫾0.01 AM vs CM,‡ AM vs AF† 4
LV⫹S, mg 79.3⫾0.3 81.3⫾8.3* 74.8⫾1.4 60.9⫾2.0 AM vs AF* 4
Blood
HCT, % 47.1⫾0.8 44.5⫾0.9 42.4⫾2.6 44.8⫾0.8 4–5
WBC, 103cells/L 2.1⫾0.5 1.4⫾0.4 2.3⫾0.6 1.2⫾0.2 4–5
Fifteen-week-old male and female mice on regular chow or HF diet for 11 weeks in normoxia. Statistically significant differences between C57Bl/6 (control) and apoE⫺/⫺mice of either gender and between genders of the same genotype are indicated. Values are mean⫾SEM. CM indicates control males; AM, apoE⫺/⫺males;
CF, control females; AF, apoE⫺/⫺females; S, septum; BP, blood pressure; MAP, mean arterial pressure; LVEDP, LV end-diastolic pressure (determined by left carotid artery/LV catheterization); EF, ejection fraction; FS, fractional shortening; CO, cardiac output; LVIDD, LV end-diastolic inner diameter; LVISD, LV end-systolic inner diameter; HCT, hematocrit; and WBC, white blood cell count.
*P⬍0.05; †P⬍0.01; ‡P⬍0.001.
period of time was not significantly different from the baseline values observed in the 15-week-old mice of the initial study (see Figure 1A compared with Figure 3B and Table 1 versus Table 2). In addition to the induction of plasma adiponectin and improvement of insulin sensitivity, a 4-week course of rosiglitazone treatment resulted in lower RVSP, RV mass (RV/LV⫹septum ratio), and percentage of muscularized arteries at the alveolar wall level that were similar to those in C57Bl/6 control mice (Figure 3B, 3D, and 3F). Because we started treatment at a time when the insulin-resistant male apoE⫺/⫺ mice already had elevated RVSP, RVH, and peripheral PA muscularization, our find-ings indicate that we had induced regression of PAH. Ros-iglitazone given for 4 weeks caused no significant differences in systemic blood pressure, heart rate, LV systolic function (fractional shortening, ejection fraction), cardiac output, and RV systolic and diastolic function (Table 2). Other measure-ments such as hematocrit and white blood cell count also were similar in treated and untreated animals. There was, however, a tendency for rosiglitazone to cause mild LV dilatation and increased LV mass in both C57Bl/6 control and apoE⫺/⫺mice (Table 2).
All the features described above were sustained after a 10-week rosiglitazone treatment period, suggesting that the effect was not transient. However, at this time point, hemat-ocrit was slightly lower in both control and apoE⫺/⫺ mice treated with rosiglitazone (Table III in the online Data Supplement). Decreased hematocrit and a tendency for LV dilatation and increased LV mass have been previously reported in the clinical setting as minor side effects of rosiglitazone treatment.37
Rosiglitazone Does Not Regulate Plasma ET-1 and ADMA in ApoEⴚ/ⴚ Mice
ET-1 and ADMA were measured in blood plasma after a 4-week treatment with rosiglitazone. Plasma ET-1 levels were lower in treated than in untreated C57Bl/6 mice. However, ET-1 levels were lower in apoE⫺/⫺compared with C57Bl/6 mice and were not affected by rosiglitazone. ADMA levels were in the normal murine range, not different between genotypes, and not altered by rosiglitazone treatment (Table 2).
Recombinant ApoE and Adiponectin Inhibit PASMC Proliferation
To support possible roles of apoE and adiponectin in protect-ing against the development of PAH, we showed that both proteins inhibit PDGF-BB–induced proliferation of PASMCs harvested from both C57Bl/6 control and apoE⫺/⫺ mice (Figure 4A and 4B).
Discussion
Although a link between insulin resistance and systemic cardiovascular disease is evident in both clinical2and exper-imental studies,3,33this report is the first indication that there may be a possible link with PAH. If insulin resistance does contribute to the pathobiology of PAH in humans, it will be an extremely important relationship owing to the steadily increasing number of children, adolescents,1and adults2with
B
E D
G F
C
Figure 1.Pulmonary hypertension in apoE⫺/⫺mice on HF diet in nor-moxia. Fifteen-week-old male mice on HF diet for 11 weeks. A, RVSP.
B, RVH, measured as ratio of the weight of the right ventricle (RV) to that of left ventricle (LV) plus septum (S). C, Muscularization of alveolar wall arteries. Bars represent mean⫾SEM (n⫽4–5 as indicated in col-umn graphs). *P⬍0.05; **P⬍0.01; and ***P⬍0.001. D and E, Repre-sentative photomicrographs of lung tissue (stained by Movat penta-chrome) of 15-week-old male mice on HF diet showing a typical nonmuscular peripheral alveolar artery in a C57B1/6 mouse (D). A similar section in the apoE⫺/⫺mouse shows an alveolar wall artery surrounded by a rim of muscle (E). F and G, Micro-CT imaging of barium-injected pulmonary arteries (PA). Representative irregularly shaped main PA wall is observed in apoE⫺/⫺mouse on HF diet (arrow), but significant PA stenoses are excluded in C57B1/6 control (F) and apoE⫺/⫺mice (G).
the metabolic syndrome, which includes insulin resistance as a key element.
ApoE⫺/⫺mice of both genders on HF showed atheroma on histology and on micro-CT, affecting the large PAs in a nonocclusive manner. Similar features have been described in pathological specimens of adult patients with PAH.38 How-ever, the cholesterol levels and the presence of atheroma were similar in male and female apoE⫺/⫺ mice, but only male apoE⫺/⫺developed insulin resistance and severe PAH (ie, a combination of marked RVSP elevation, RVH, and enhanced peripheral PA muscularization). We therefore suggest that apoE deficiency, hypoadiponectinemia, and the related insu-lin resistance, rather than pulmonary atherosclerosis itself, were the major causes of PAH.
The level of PAH seen in our model, with an RVSP baseline elevation of 7 to 9 mm Hg over controls for males on a HF diet in normoxia, was not based on LV dysfunction and is comparable to39 or even greater than40,41 that in the relatively few murine models of PAH described to date. The sole exception is the inducible vascular SMC dominant-negative bone-morphogenetic protein receptor II (BMP-RII) transgenic mouse. However, values in the control mice of this
BMP-RII model also were quite elevated, consistent with the relative “hypoxia” at Denver altitude.42
The likely mechanism by which apoE deficiency leads to the development of PAH appears to be facilitated by PDGF-BB signaling. From studies in systemic vascular SMCs, we know that PDGF-BB signaling in SMCs is suppressed when apoE binds to the low-density lipoprotein receptor–related protein (LRP), thereby initiating endocytosis and degradation of the LRP–PDGFR-–PDGF-BB com-plex.28 –30 We have shown here that PDGF-BB–induced proliferation also is suppressed by apoE in PASMCs, sug-gesting that a similar mechanism may be present in the pulmonary vasculature. In diabetic apoE⫺/⫺mice, PDGFR- signaling is increased in vascular SMCs, and systemic vas-cular disease can be reversed by the PDGFR tyrosine kinase inhibitor imatinib.32 We40 and others21 have shown that blockade of tyrosine kinase activity selective to the epidermal growth factor receptor40or the PDGFR21reverses experimen-tal PAH and vascular remodeling caused by the toxin mono-crotaline in rats.
In association with insulin resistance, the male mice (apoE⫺/⫺and C57Bl/6) had lower levels of adiponectin than
A B
E F
C D
Figure 2.Male but not female apoE⫺/⫺mice develop insulin resistance and hypoadiponectinemia on HF diet. Plasma adiponectin (A,B), blood glucose (C, D), and plasma insulin (E, F). Overnight-starved 15-week-old male (A, C, E) and female (B, D, F) C57Bl/6 and apoE⫺/⫺
mice, on regular chow or HF diet for 11 weeks. Note that C57Bl/6 mice but not apoE⫺/⫺mice of both genders upregulate their adi-ponectin levels when exposed to HF diet. Bars represent mean⫾SEM (n⫽5–10 as indicated in column graphs). *P⬍0.05; **P⬍0.01; and
***P⬍0.001.
the female mice. This finding is in keeping with a recent study demonstrating that testosterone inhibits the secretion of adiponectin in adipocytes.35 Adiponectin reverses insulin resistance26 and is independently associated with a reduced risk of type 2 diabetes mellitus in apparently healthy individ-uals.43 Moreover, the high-molecular-weight form of adi-ponectin binds PDGF-BB, thereby reducing PDGF-BB bio-availability31and mitogenic postreceptor function in vascular SMCs.27Adiponectin has been shown to be a transcriptional target of PPAR␥ in adipocytes.7 Because previous studies used the PPAR␥ agonist rosiglitazone to suppress intimal thickening in the apoE⫺/⫺mouse,44we reasoned that it might also be effective in preventing disease progression or revers-ing PAH. Indeed, treatment of male apoE⫺/⫺ mice with rosiglitazone increased plasma adiponectin, improved insulin sensitivity, and led to sustained regression of PAH, RVH, and abnormal muscularization of distal PAs. The source of adiponectin in our animal model is likely from visceral and subcutaneous as well as perivascular adipocytes.17To further support a mechanistic relationship between apoE and adi-ponectin deficiency on the one hand and PAH with enhanced muscularization of peripheral arteries on the other, we showed that both recombinant apoE and adiponectin inhibited
PDGF-BB–induced proliferation of cultured murine wild-type and apoE⫺/⫺PASMCs.
In addition to elevating adiponectin levels and thus causing sequestration of PDGF-BB, PPAR␥activation blocks PDGF gene expression45 and PDGF-BB–mediated systemic SMC proliferation and migration.33,34 This is likely to occur via inhibition of phosphorylated extracellular-regulated kinase nuclear translocation46and/or induction of protein phospha-tases47 that reduce phosphorylated extracellular-regulated kinase. Furthermore, PPAR␥ induces expression of LRP,48 the receptor necessary for apoE-mediated suppression of PDGF-BB signaling.28 –30 By blocking important survival pathways downstream of activated PDGFR- (eg, PI3K),25 rosiglitazone could also induce apoptosis of proliferating vascular cells.33,49
We considered the possibility that PPAR␥activation might impair the expression of ET-115 and the endogenous nitric oxide synthase inhibitor ADMA,18 both of which have previously been linked to clinical PAH16,19 and insulin resistance.17,18 However, although it is not surprising that rosiglitazone decreased ET-1 levels in the C57Bl/6 mice, we expected the apoE⫺/⫺ mice to have higher levels of ET-1 under control conditions. In the present study, however, we
C D
E F
Figure 3.Four-week treatment with the PPAR␥agonist rosiglitazone reverses PAH, increases plasma adiponectin, and induces insulin sensitivity. Measurements of plasma adiponectin (A), blood glucose (C) and plasma insulin (E), RVSP (B), RVH (D), and muscularization of alveolar wall arteries (F). Nineteen-week-old male C57Bl/6 and apoE⫺/⫺mice, all on HF diet for 15 weeks, were used. Bars represent mean⫾SEM (n⫽4 –16 as indicated in column graphs). *P⬍0.05; **P⬍0.01; and ***P⬍0.001.